Embed Size (px)
Citation preview
FACTORS AFFECTING THE BREAKING OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
by
SIDNEY JOHN PEARCE
A THESIS
submitte-d to
OREGON STATE COLLEGE
in partial fulfillment of the requirements for the
degree of
MASTER OF SCIENCE
June 1953
rrPnamptDr
trtfcmw of ampiry Amaaf Xtr ghrtga cf IrJ
Redacted for Privacy
Redacted for Privacy
Redacted for Privacyfurmn af SchrI diqrmtr Gmlttm
Eco of Eredutto Schoot
Redacted for Privacy
Drtr thrtr lr prcrcatre Sprd bf Jrnt 3 BEral
ACKNOWLEDGMENTS
I would like to thank Dr G A Richardson for his
guidance over the research and helpf-ul criticism or the
manuscript
I would also like to express my gratitude to
Dr Jerome C R L1 for his help in directing the
statistical analyses
Finally appreciation is expressed to the Robm and
Haas Chemical Corporation Philadelphia Pennsylvania
the Wyandotte Chemical Corporation Wyandotte Mi chigan
and the Rumford Chemical Works Rumford Rhode Island
for the samples of Triton x-100 Pluronics and sodium
tetraphosphate respectively which were used in this
study
bull bull bull bull bull
bull bull
TABLE OF COl TENTS
Page
INTRODUCTION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 1
REVIEW OF LITERATURE bull bull bull -middot bull bull bull bull bull bull bull bull bull bull bull 3
EXPERI ffiNTAL bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10
Babcock Test bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Roese-Gottlieb Method bull bull bull bull bull bull bull bull bull bull bull bull 10 Detergent Analyses bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Modifications bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 11 Reagents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 12 Pluronic Surface Active Agents bull bull bull bull bull bull bull bull 12
STATISTICAL ANALYSIS OF EXPERif~fTAL DATA 13
Comparison of Babcock BDI Detergent ethods by Jersey and Holstein Breeds and Cows within each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 13
Comparison of BDI Babeoek and Roese-Gottlieb (Mojonnier) with Homogenizedmiddot Milk bull 15
Effect of Variation of Triton X-100 and Sodium Tetraphosphate Concentrations upon the Accuracy of the Test bull bull bull bull bull bull bull bull bull bull 17
Stability of Reagent upon Storage and Its Relation to Accuracy bull bull bull bull bull bull bull bull bull _ bull 21
Rate and Completeness of Fat Liberation without Centrifuging bull bull bull bull bull bull bull bull bull bull bull bull 21
GENERAL PHYSICAL AND CHEMICAL PHENOMENA CONCERNING DE- EIIDLSIFICATION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 24
Studies with Washed Cream bull bull bull bull bull bull bull bull bull bull 25 SUrface Tension Studies bull bull bull bull bull bull bull bull bull bull 25 Viscosity Ghanges in the Milk-Detergent
lUxtures bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 28 Nature of the Tetraphosphate Precipitate bull bull bull 57 Action of Methyl Alcohol bull bull bull bull bull bull bull bull bull 45 Use of Dye to Observe Progress of
De-emulsification bull bull bull bull bull bull bull bull bull bull bull bull bull 46 Studies Involving Pluronie Surface-Active
Agents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 49
DISCUSSION bull bull bull bull bull bull bull bull bull bull bull bull bull bull 52bull bull bull li
Stati stical Analyses bull bull bull bull bull bull bull bull bull 52
TABLE OF CONTENTS (CONTINUED)
Page
General Phenomenon Related to Detergent Action 55
SUJfitiARY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 63
BIBLIOGRAPHY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 66
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
rrPnamptDr
trtfcmw of ampiry Amaaf Xtr ghrtga cf IrJ
Redacted for Privacy
Redacted for Privacy
Redacted for Privacyfurmn af SchrI diqrmtr Gmlttm
Eco of Eredutto Schoot
Redacted for Privacy
Drtr thrtr lr prcrcatre Sprd bf Jrnt 3 BEral
ACKNOWLEDGMENTS
I would like to thank Dr G A Richardson for his
guidance over the research and helpf-ul criticism or the
manuscript
I would also like to express my gratitude to
Dr Jerome C R L1 for his help in directing the
statistical analyses
Finally appreciation is expressed to the Robm and
Haas Chemical Corporation Philadelphia Pennsylvania
the Wyandotte Chemical Corporation Wyandotte Mi chigan
and the Rumford Chemical Works Rumford Rhode Island
for the samples of Triton x-100 Pluronics and sodium
tetraphosphate respectively which were used in this
study
bull bull bull bull bull
bull bull
TABLE OF COl TENTS
Page
INTRODUCTION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 1
REVIEW OF LITERATURE bull bull bull -middot bull bull bull bull bull bull bull bull bull bull bull 3
EXPERI ffiNTAL bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10
Babcock Test bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Roese-Gottlieb Method bull bull bull bull bull bull bull bull bull bull bull bull 10 Detergent Analyses bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Modifications bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 11 Reagents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 12 Pluronic Surface Active Agents bull bull bull bull bull bull bull bull 12
STATISTICAL ANALYSIS OF EXPERif~fTAL DATA 13
Comparison of Babcock BDI Detergent ethods by Jersey and Holstein Breeds and Cows within each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 13
Comparison of BDI Babeoek and Roese-Gottlieb (Mojonnier) with Homogenizedmiddot Milk bull 15
Effect of Variation of Triton X-100 and Sodium Tetraphosphate Concentrations upon the Accuracy of the Test bull bull bull bull bull bull bull bull bull bull 17
Stability of Reagent upon Storage and Its Relation to Accuracy bull bull bull bull bull bull bull bull bull _ bull 21
Rate and Completeness of Fat Liberation without Centrifuging bull bull bull bull bull bull bull bull bull bull bull bull 21
GENERAL PHYSICAL AND CHEMICAL PHENOMENA CONCERNING DE- EIIDLSIFICATION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 24
Studies with Washed Cream bull bull bull bull bull bull bull bull bull bull 25 SUrface Tension Studies bull bull bull bull bull bull bull bull bull bull 25 Viscosity Ghanges in the Milk-Detergent
lUxtures bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 28 Nature of the Tetraphosphate Precipitate bull bull bull 57 Action of Methyl Alcohol bull bull bull bull bull bull bull bull bull 45 Use of Dye to Observe Progress of
De-emulsification bull bull bull bull bull bull bull bull bull bull bull bull bull 46 Studies Involving Pluronie Surface-Active
Agents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 49
DISCUSSION bull bull bull bull bull bull bull bull bull bull bull bull bull bull 52bull bull bull li
Stati stical Analyses bull bull bull bull bull bull bull bull bull 52
TABLE OF CONTENTS (CONTINUED)
Page
General Phenomenon Related to Detergent Action 55
SUJfitiARY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 63
BIBLIOGRAPHY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 66
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
ACKNOWLEDGMENTS
I would like to thank Dr G A Richardson for his
guidance over the research and helpf-ul criticism or the
manuscript
I would also like to express my gratitude to
Dr Jerome C R L1 for his help in directing the
statistical analyses
Finally appreciation is expressed to the Robm and
Haas Chemical Corporation Philadelphia Pennsylvania
the Wyandotte Chemical Corporation Wyandotte Mi chigan
and the Rumford Chemical Works Rumford Rhode Island
for the samples of Triton x-100 Pluronics and sodium
tetraphosphate respectively which were used in this
study
bull bull bull bull bull
bull bull
TABLE OF COl TENTS
Page
INTRODUCTION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 1
REVIEW OF LITERATURE bull bull bull -middot bull bull bull bull bull bull bull bull bull bull bull 3
EXPERI ffiNTAL bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10
Babcock Test bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Roese-Gottlieb Method bull bull bull bull bull bull bull bull bull bull bull bull 10 Detergent Analyses bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Modifications bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 11 Reagents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 12 Pluronic Surface Active Agents bull bull bull bull bull bull bull bull 12
STATISTICAL ANALYSIS OF EXPERif~fTAL DATA 13
Comparison of Babcock BDI Detergent ethods by Jersey and Holstein Breeds and Cows within each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 13
Comparison of BDI Babeoek and Roese-Gottlieb (Mojonnier) with Homogenizedmiddot Milk bull 15
Effect of Variation of Triton X-100 and Sodium Tetraphosphate Concentrations upon the Accuracy of the Test bull bull bull bull bull bull bull bull bull bull 17
Stability of Reagent upon Storage and Its Relation to Accuracy bull bull bull bull bull bull bull bull bull _ bull 21
Rate and Completeness of Fat Liberation without Centrifuging bull bull bull bull bull bull bull bull bull bull bull bull 21
GENERAL PHYSICAL AND CHEMICAL PHENOMENA CONCERNING DE- EIIDLSIFICATION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 24
Studies with Washed Cream bull bull bull bull bull bull bull bull bull bull 25 SUrface Tension Studies bull bull bull bull bull bull bull bull bull bull 25 Viscosity Ghanges in the Milk-Detergent
lUxtures bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 28 Nature of the Tetraphosphate Precipitate bull bull bull 57 Action of Methyl Alcohol bull bull bull bull bull bull bull bull bull 45 Use of Dye to Observe Progress of
De-emulsification bull bull bull bull bull bull bull bull bull bull bull bull bull 46 Studies Involving Pluronie Surface-Active
Agents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 49
DISCUSSION bull bull bull bull bull bull bull bull bull bull bull bull bull bull 52bull bull bull li
Stati stical Analyses bull bull bull bull bull bull bull bull bull 52
TABLE OF CONTENTS (CONTINUED)
Page
General Phenomenon Related to Detergent Action 55
SUJfitiARY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 63
BIBLIOGRAPHY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 66
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
bull bull bull bull bull
bull bull
TABLE OF COl TENTS
Page
INTRODUCTION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 1
REVIEW OF LITERATURE bull bull bull -middot bull bull bull bull bull bull bull bull bull bull bull 3
EXPERI ffiNTAL bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10
Babcock Test bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Roese-Gottlieb Method bull bull bull bull bull bull bull bull bull bull bull bull 10 Detergent Analyses bull bull bull bull bull bull bull bull bull bull bull bull bull 10 Modifications bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 11 Reagents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 12 Pluronic Surface Active Agents bull bull bull bull bull bull bull bull 12
STATISTICAL ANALYSIS OF EXPERif~fTAL DATA 13
Comparison of Babcock BDI Detergent ethods by Jersey and Holstein Breeds and Cows within each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 13
Comparison of BDI Babeoek and Roese-Gottlieb (Mojonnier) with Homogenizedmiddot Milk bull 15
Effect of Variation of Triton X-100 and Sodium Tetraphosphate Concentrations upon the Accuracy of the Test bull bull bull bull bull bull bull bull bull bull 17
Stability of Reagent upon Storage and Its Relation to Accuracy bull bull bull bull bull bull bull bull bull _ bull 21
Rate and Completeness of Fat Liberation without Centrifuging bull bull bull bull bull bull bull bull bull bull bull bull 21
GENERAL PHYSICAL AND CHEMICAL PHENOMENA CONCERNING DE- EIIDLSIFICATION bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 24
Studies with Washed Cream bull bull bull bull bull bull bull bull bull bull 25 SUrface Tension Studies bull bull bull bull bull bull bull bull bull bull 25 Viscosity Ghanges in the Milk-Detergent
lUxtures bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 28 Nature of the Tetraphosphate Precipitate bull bull bull 57 Action of Methyl Alcohol bull bull bull bull bull bull bull bull bull 45 Use of Dye to Observe Progress of
De-emulsification bull bull bull bull bull bull bull bull bull bull bull bull bull 46 Studies Involving Pluronie Surface-Active
Agents bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 49
DISCUSSION bull bull bull bull bull bull bull bull bull bull bull bull bull bull 52bull bull bull li
Stati stical Analyses bull bull bull bull bull bull bull bull bull 52
TABLE OF CONTENTS (CONTINUED)
Page
General Phenomenon Related to Detergent Action 55
SUJfitiARY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 63
BIBLIOGRAPHY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 66
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
TABLE OF CONTENTS (CONTINUED)
Page
General Phenomenon Related to Detergent Action 55
SUJfitiARY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 63
BIBLIOGRAPHY bull bull bull bull bull bull bull bull bull bull bull bull bull bull bull 66
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
bull bull
LIST OF TABLES
Table Page
I Comparison of BDI Method with Babcock According to Breed and Cows within each Breed bull middot bull bull ~ bullmiddot bull bull bull bull bull bull bull bull bull 14
II Comparison ot BDl middotBabcock and Roese-Gottlieb MoJonnier) with Homogenized Milk Reagent Stability bull bull bull bull bull bull bull bull bull bull 16
III Effect of Variation in Triton x-100 and Sodium Tetztaphosphate Concentration upon the Test Accuracy bull bull bull bull bull ~ ~ bull bull bull bull 18
IV Reagent Concentmiddotrations imiddotor Analysis of Effecmiddott of Varyfng Leve1s of Sodium Tetraphosphate and Triton X-100middot on Fat Liberation bull bull bull bull bull bull bull bull bullmiddot bull bull 19
V Stability of Reagent-s bullbullbullbullbullbull 22
VI Observed Rate of Fat Rise without Centrishyfug~ng Sample bull bull bull bull bull bull bull bull bull bull bull bull 23
VII SUrface Tension Measurements on Babcoek and Detettgent 8ystems 370 C bull bull bull bull bull bull bull 29
VIII Viscosity DElterminations on Various Detergent- Milk ~ystems bull bull bull bull bull bull bull bull bull bull 33
IX Relationship Between Protein Distribution and Precipitate Formation with Variation in Polyphoaphate Concentration - bull ~ bull bull 42
X Fat Test Values for Comparison of BDl Method with Babcock according to Breed and Cowa Within Each Breed bull bull bull bull bull bull bull bull bull bull bull bull bull 69
XI Fat Test Values for Analysis of Effect of Varying Levels of Sodium Tetraphosphateand Triton X-100 on Fat Liberation bullbull 70
XII Test Values f or Comparison of BDI Babcock and Roese-Gottlieb (Mojonnier) with Ho-mogenized Milk bull bull bull bull bull bull bull bull bull bull 71
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
bull bull bull bull bull bull bull
LIST OF FIGURES
Figure Page
1 Variation in Na Tetraphosphate 20
2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns 27
3 Influence of Na Tetraphosphate on Viscosity 31
4 The Effect of Variation of Sodium Tetrashyphosphate upon Protein Hydration in Detemiddotrgent Tests bull bull bull bull bull bull bull bull bull bull bull bull bull 39
5 Inorganic Precipitate Formation with Exshycesses of Sodium Tetraphosphate in Milk 40
6 Comparison of Detergent Tests in which 50$ Methanol (left) and Distilled Water (right) were used to Float the Fat into the Neck of the Bottle bull bull bull bull bull bull bullbull 41
7 The Adsorption and De-adsorption of Methylene- blue Dye from the Surface of the Detergent Test during Fat Liberation bull bull bull bull 48
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
FACTORS AFFECTING THE BREAKIUG OF MILK AND CREAM EMULSIONS BY SURFACE-ACTIVE AGENTS
INTRODUCTION
The advent of synthetic detergents and surface-active
agents has brought with it many innovations and time-saving
practices to the field of food-plant sanitation Only
recently work first initiated by Schain and lately culmishy
nated by Sager and co-workers at the USDA has indicated
that these compounds have the potential of supplanting the
long established but inconvenient Babcock acid-test for
milk fat in dairy products
Wbile it has been demonstrated that these surface-
active agents are capable or breaking the milk fat
emulsion there has been as yet no statistical evidence of
the quantitative efficiency of any of these methods so far
devised Since the Bureau of Dairy Industrys method BDI)
appears to be the most promising it constituted the basis
of most of this research It was first studied as outlined
by the USDA and then modified in its construction and
performance to provide same insight into the test mechanism
itself
In addition to the nonionic detergent of the BDI
technique other surface-active agents were studied to
compare modes of action
The principle of the detergent test was studied with
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
2
the main objective ot demonstrating whether or not these
materisls could p~ov1de a method of quant-itative rat
determina tion with the ease simplicity$ and precision ef
the Babcock method If these conditions could be met
the great advantage of the non-corroaiveness of the
detergent would place a safe valuable tool into the hands
of the average milk tester
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
3
REVIEW OF LITERATURE
The first observed attempt to apply synthetic detershy
gents to de-emulsification of the fat in milk and cream
was made by Schain in 1949 (20 ppl21-122)(21 pp 29 50
54 56 He used two surface-active ents detergents)
which were anionic dioctyl sodium phosphate (Tergitol
P-28) and nonionic polyoxyethylene sorbitanmonolaurato
(Tween 20) as the basis of a quantitative fat test for
milk By first mixing the anionic reagent with the sample
and then carefully layering the nonionic material on the
bottom and heating at 80deg c for a few minutes in a water
bath he was able to report the test on 100 duplicate
milks to be the same as respective Babcock values He
described the reaction mechanism as functioning through
the formation of a protein-detergent complex consisting of
the anionic detergent and the protein film on the fat
globule Accordingly after the film or globule membrane
was dispersed the fat was liberated thereby allowing it
to coalesce with othe r fat globules Sehain remarked
however that the separation at this point tas not complete
but uwhen the strongly hydrophilic nonionic detergent is
added a clear solution and complete separation results
(20 pl22)
In later work Schain (22 pp l2-17) substituted
tetradecyl desoxy-polyethylene glycol for the nonionic
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
4
Tween 20 of his original method and modified his technique
making it somewhat more delicate At this time however
he found it necessary to employ a nomograph to correct
for deviations in observed values from the Roese-Gottlieb )
He ascribed three factors (a) speeiric gravity (b)
separate layering of reagent and milk and (c) selective
application of heat the responsibility for proper detershy
gent action on milk fat This action was considered to be
an extraction11 of the fat by the detergent occurring as
the selective heating of the Babcock bottle caused an
interchange in the relative positions of the two carefully
layered materials For two series of tests involving
fifty duplicate samples Schain reported that the average
deviation from Roese-Gottlieb for the corrected detergent
tests was less than the average deviation of either Babcock
or Gerber tests of the same samples
Gres chenfeld and Ucko (4 pp l75-l76) modified the
original Schain technique with respect to detergent conshy
centrations and heating procedure In each test they
found that the amount of nonionic detergent used must first
be standardized to an optimum quantity to produce a test
equal to the Babcock value Homogenized milk was found to
require more nonionic than did regular milk and milk of a
fat content greater than 4 per cent required a smaller
quantity of nonionic detergent to give results equivalent
to t he Babcock test
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
5
The seope of this test was broadenGd (5 pp342-343)
to include other dairy produets such as butter and eheese
No mention was made however of any necessity of
standardizing the amounts of detergent solutions employed
according to the nature of the product or weights to be
tested
The latest vork by Greshenfeld with Rosenthal (6
ppl7-18) described techniques in which they employed
several other anionic and nonionic detergents--1veen
(20 40 80 and 85) Tritons and Tergitols In their
milk fat determinations perfect agreement with Babcock
readings was reported in every case
Patton and Stein ( 14 ppbull324 528) (23 pp Bffi- 660)
studied a wide variety of organic compounds and surface
active agents to ~termine those which might be effeetively
used to de-emulsiry cream in the preparation of butter-oil
Among the group of 69 organie compounds studied they
observed a combination of n-butylamine and n-butanol to be
the most effective and of the 97 surface-active agents
studied 26 were found capable of de-emulsifying cream
~venty~four ef these surface-active materials were anionic
and two were nonionic Tergitol-7 was found to be the most
effective member of this group
More detailed work with Tergitol-7 disclosed that t~
compound is effective in quantitatively de-emulsifying
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
6
cream when used at levels of 10 per cont or less at an
optimum temperature of 80deg c for approximately fifteen
minutesmiddot Analysis for sulfur indicated that no appreciable
quantity of detergent had dissolved in the fat ~ This as
further substantiated by the absence of flavor changes
upon re-emulsification of the treated rat ~n skim milk
Wildasin (25 ppbull 87- 92) (26 pp l-10) added cationic
detergents to the acid of the prosent Babcock test to
provide for more complete fat liberation especially in
the case of homogenized fluid milk products The authors
asserted that the acid-hydrolyzed proteins are rendered
more soluhle by tho quaternary ammonium compounds with the
result that fewer burned or curdy particles appeared in
the fat column
Sager and co-workers 18 pp 27-43) compared the
Schain test ith the Roese- Gottlieb and the modified
Schain and found a non-consistent deviation which makes
the application of nomograph hazardous They believed the
deviation of observed values from the official test is due
to (a) the entrapment of f at globules in the viscous
protein complex formed with the anionic detergent and by
(b) the solubility of the nonionic detergent in the fat
The latter authors observed that some of the detershy
gents studied (Triton series) 1ere precipitated from
solution when heated and upon the addi tion of salt
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
7
solution~ they floated on the surface On the basis of
these observations several detergents were classified
according to the volume of saturated sodium chloride
required to precipitate them (cloud point) at 6QO C This
served to measure the hydrophilic attraction of the detershy
gent molecule That the anionic detergents as a class
required the greatest volume of saturated salt to reach
the cloud point justified the authors theoretical
contentions Incidentally it was found that those detershy
gents requiring only a small amount of salt to be made to
cause them to float on the surface were oil soluble The
Schain reagent is an example
A quantitative test for fat in milk using Triton
X-100 (a conjugation product of ethylene oxide and dishy
isobutylphenol) and sodium chloride was developed by these
authors--Sager and eo-workers-who explained the fat
liberation mechanism as a detergent salting-out effect
They considered that while the detergent is associated
with the fat globule there remained a strong hydrophilic
attraction which can be broken upon the application of
heat 1n the presence of the salt medium This allows the
fat-detergant association to float When the 50 per cent
methyl alcohol is added to raise the fat into the neck of
the bottle the resultant dilution shifts the affinity of
the detergent from the fat complex it has formed to the
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
8
aqueous phase ihose detergents with the low~st salt
titration values were found to be the most effective deshy
emulsifiers It tJas also found that use of too mueh salt
in the fat test resulted in incomplete fat separation
Further research with Sanders (19 pp l-14) demonshy
st~ated that the salt could be effectively substituted for
by certain members of the polymerized phosphate group
The one found most suitable was sodium tetraphosphate It
was selected upon the basis of stability and reduced hygroshy
scopic tendency compared to the hexameta and heptaphosshy
phates and themiddot greater ability to dispelsa the milk
proteins and ealcium salts than the lower members of the
phosphate series The level ot sodium tetraphosphate ta
be used 035 g 18 g milk) was determined by the salt
titration technique of their first paper A 50 per cent
methyl a~cohol solution was added upon completion opound fat
liberation to bring the fat column up to the graduated
portion of the test bottle and at the same middotcime to
clarify the fat column of eurd11 or other material intershy
mixed or associated with the fat It was found possible to
obtain a satisfactory test without centrifuging although
use of the eentrifuge was recommended
fhe rew observations made with this moctlfi cation of
the detergent test by the Bureau of Dairy Industry USDA
indicated good agreement with Babcock and Roese- Gottlieb
tests One of the distinguishing features of this test
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
eompaled to other detergent fat tests 1s its greater
s1mpl1e1ty oi operation
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
10
EXPERIMENTAL
Methods and Materials
Babcock Test This test was per~ormed according to
the official methods o~ tlw Association of Official
Agri cultural Chemists (2 pp bull 233-234)
Roese-Gottlieb Method Fat determinations by the
ether-extraction procedure were performed with Mojonnier
equipment according to the method as outlined by the
Mojonnier Company (11 pp26-27) This involves weighing
a 10 g sample denaturing the -emulsion with ammonium
hydroxide and ethanol and extracting fat with ethyl and
petrolium ether The extracted fat is then weighed in an
aluminum dish after nppropriate drying
Detergent Analyses Unless otherwise modified the
dotergent analyses were per~ormed in conformity with the
Bureau of Dairy Industrys (BDI) method and ware conducted
according to the ~ollowing procedure for milk (19 pp3-4)
The test bottle and pipette are the same as in the official Babcock test Methods of the AOAC ~Likewise the centrifuge calipers and water bath for tempering the test are the same The milk is prepared as directed on page 227 or the methods 1hich speci~ies bringshying the milk to 20deg c be~ore sampling The test is then conducted as follows
Part A with a pipette trans~er 18 g of prepared
sample to milk test bottle Blow out milk in pipette tip after free over~low has ceased
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
11
Add so ml of reagent A portionwise so as to wash all traces of milk into bulb Shake to mix Transfer bottle to bath of boiling water level of water covering level if milk is in bottle After approximately 5 minutes in bath shake to remix raised cream and replace for 10 more minutes Remove from bath Do not remix contents
Part B ieEhod using centrifuge Yfrdle bottle
is still hot add so methyl alcohol to topof graduated scale (No water is added in this procedure ) Allo alcohol to run down side of neck Transfer bottle to unheated centrifuge Rotate centrifuge 2 minutes Transfer bottle to warm water bath maintainshyed at 55-60 C immerse it to level of topof fat column and leave until column is in equilibrium and lower fat surface has assumed final form about 15 minutes at least
Method not using centrifuge While bottle is stirr-hot add so methyl alcohol pouring alcohol down side of neck until contents of bottle reach into neck of bottle but not above the zero graduation on the scale Add water until contents reach top of calibrations bullbullbullTransfer bottle to water bath at 55-60 C for tempering and continue as in Part B
Modifications The test procedure was often modishy
fied by one or more of the following changes which will
be referred to in greater detail in the sections in which
they are applied (a) Use of a five instead of a two
minute period for centrifuging b) addition of about l m1
50 per cent methanol to the test bottle in water bath pound1ve
minutes prior to end of heating period (c) addition of a
water-soluble dye (methylene blue) to the detergent solushy
tion to aid detection of unreacted cream particles (cream
which had not been de-emulsified) and (d) reading fat
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
12
test at the end of five minutes instead of a 15-minute
period in the tempering bath
Reggents The detergent solution used in the BDI
method is prepared with nonionic Triton X-100 which is a
condensation product or diisobutylphenol and ethylene
oxide It has a molecular weight of 600 and contains an
average of nine or ten ethylene oxide groups per mole
This material is used at a concentration of 3 per cent
The polyphosphste is sodium t etraphosphate Na5P4013 at a
concentration of 7 per cent This polyphosphate has a
slightly basic reaction (pH of a 1 so~ution 79) and is
highly soluble The solution of Triton X-100 and sodium
tetraphosphate was pFepared weekly although there was no
observed separation or other indications of deterioration
at room temperature well after this time
Pluronic Surpoundaee Active Agents The pluronic surfac~
active agents studied are condensation products of ethylene
oxide propylene and propylene glycol The members of
the series studied were Pluronic L-64 L-62 and F-68
Their detailed description appears in the section devoted
to the studies of the relationships of their structural
characteristics to the efficiency of de-emulsification
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
13
STATISTICAL ANALYSIS OF EXPERIMENTAL DATA
pomparison of Babcock BDI Detergent Methods ~
Jersey and Holstein Breeds and Covs within each Breed In
this first study of the accuracy of the detergent method of
the Bureau of Dairy Industry USDA as proposed by Sanders
and Sager (19 pp 3-4) 12 samples each of Jersey and
Holstein evening milkings were obtained The method of
the Bureau of Dairy Industry for the detergent test was
followed except that a five-minute period was used for
centrifuging samples and readings were made aftor the usual
fi ve-minute tamporing period Babcock analyses were pershy
iormed according to official methods AOAC (2 pp 233-234)
Three rep11cations of each of the 24 samples by both
methods were made 1n order that a comparison of the two
methods on milks of varying fat contents within each breed
could be determined The results of the analysis of
variance are presented in Table I
It is observed that at the 5 per cent level of
significance there is no quantitative difference between
the observed fat content of any individual sample of either
breed as determined by the two tests The standard deviashy
tion in this analysis was 013 which is believed to be
acceptable for Babcock analysis
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
I
14
TABLE I
COMPARISON OF BDI tETHOD WITH BABCOCK ACCORDING TO BREED ~~D C01S
WITHIN EACH BREED
Analysis of Variance
Variation SUm or
Deg of Free- Mean
Due to Squares dom Square F Remarks
Bre-ed 205 46833 l 20546833 ll 781 44 Sig
Cow 10738941 22 4 88134 27989 Sig
ethod 000855 1 0 00855 0487 Not sig
Math X Brd 000059 1 oooose 0 003 Not sig
Math x Cow 0 47391 22 0 02154 123 Not sig5
Replication 0 01218 2 0 00609 4 08
Rep x Brd 0 02986 2 0 00149 0752 Not sig
Rep x Cow 087200 44 000198 0094 Not sig
Rep x Math 004202 2 q02101 215 Not sig
Rep X Math X Brd 001947 2 0 00979 0593 Not sig
Rep X Meth X cow 069301 42 0 01650
Total 31500933 141 0 22341
Pooled Error 94 0 017442
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
15
Comparison of BDI Babcock ~ Roese-Gottlieb
(Mojonnier) ~ Homogenized ~middot It is presently
recognized that the official Babcock procedure is not
entirely satisfactory middotas applied to homogenized milk
Common complaints are curdy fat columns burned particles
in fat column and low values
This experime11t combined an analysis of the accuracy
of the Babcock and BDI technique relative to the Mojonnier
ether extraction with a deterwination of the efpoundeet of
age upon the activity of the detergent reagent This was
achieved by testing six samples of the BDI reagent of
varying ages including _one sample which had been held at
36deg C for two weeks in addition to its having aged at room
temperature The BDI method was followed as indicated in
the previous experiment with the added modification of
the addition of about 1 ml of 50 per cent methanol to the
reaction mixture i n the Babcock bottle five minutes prior
to the end of the heating period This merely served to
aid in clarifying the fat by acting on any unreacted crenm
during the final stages of heating The Babcock test was
conducted as indicatedabove the acid being added in three
installments Six replications of each reagent or test
method were obtained using commercially homogenized milk
The statistical analysis found in Table II
demonstrates that as in the first experiment there was no
significant difference between replications There was
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
TABLE II
COMPARISON OF BDI BABCOCK AND ROESE-GOTTLIEB (MOJOJlliERS WITH HOMOGENIZED MILK
REAGEUT STABILITY
Analysis of Vallianoe
sum Deg of Variation or Free- Mean Due to Squares dom Square
Replication oo3l894 5 0006379 Reagent 15middot24698 7 0217814 Error 0120689 35 0~003448 Total 1677281 47
1~6 1 257358 5 0251472 7 vs apound 0163333 1 0163333 (1-6) vs(748) 0104006 1 0104006
~t-Detei-gent reagents middotsee Table V for ages
~Babcock vs Mojonnier
Tukeys LSD middotReagent Means
Reagent (Method) Means
2 - BDI 17 days old 42550 3 u1 - BDI 4 bull 2266
3 BDI 25 ft ~2116 u u4 - BDI 58 4~1183 Mojonnier 4_0967 5 middot~ BDI 64 days middotold 3 9967 Babcock 38633 6 - BD-I 126 days old 3~7167
F Remarks
185 not sig 6317 Sig
-7293 4737 3016
Difference
LSD 11 0892 bull000~0289
00150 00933 o ~Oll6 o ioo 0 bull 334 0-0466
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
17
however a significant difference in results as obtained
by the three different methods of analysis Breakdown of
the treat1nent effect into its component parts followed by
analysis on the basis of single degrees of freedom ilshy
luminates these differences First it is found there is
a significantly lower value for Babcock analysis compared
to tojonnier Secondly there is observed a significant
difference among results obtained by detergent reagents of
various ages
Grouping of sample means by Tukeys method (24 pp
99-114) showed that only reagent No 4 which was 58 days
old could be considered to give a fat test the same as
could be obtained by ether extraction The most recently
prepared re ants showed mean values significantly higher
than MoJonnier whereas bcoek and detergent reagents No
5 and 6 (which were considerably old see Table III)
produced significantly lower results
Effect of Variation of Triton X-100 ~ Sodium
Tetraphosphate Concentrations upon EEl Accuracy 2pound 2
~middot To demonstrate the relative effect which different
levels of the two main reagents would have on the quantity
of fat released a series of eight replications of 24 difshy
ferent combinations of the two reagents were prepared as
indicated in Table III The results of the analysis
variance (Table IV) indicate that the effect of variation
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
18
TABLE III
EFFECT OF VARIATIO l HI TRITOli X-100 AUD SODIUM TETRAPHOSPHATE COUCENTRATION UPON THE TEST ACCURACY
Analysis of Variance
Degrees Variation Sum of of Mean Due to Squares Freedom Square F
Replication 0582062 7 0083152 276 s
Detergent 557715173 5 111543035 1505 s Tetra 21658639 3 7219546 097 NS
Det x Tetra 111182583 15 7412172 24596
Error 4851826 161 0030136
Total 695990283 191
in ~he sodium tetraphosphate level used was not significant
although variation in the concentration of the detergent
did cause a significant difference in the quantity or fat
liberated Howsectver when the effeet or zero level of
detergent was removed from the detergent treatment sum of
squares through decomposition and analysis of single
degrees of freedom it was found that there was no signifishy
cant difference in the effect upon fat readings resulting
from variation of the deter gent levels indicated Figure 1
shows the appeat~ance of detergent test following reaction
with the four levels of sodium tetraphosphate
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
19
TABLE IV
REAGENT CONCENTRATIONS FOR ANAIXSS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHATE
AND TRITON X-100 ON FAT LIBERATION
Sample ~ Na Tetra- Triton Sample ~ Na Tetra- Triton No phosphate x-1oo No phosphate X-100
1 0 0 13 194 0
2 0 056 14 194 056
3 0 111 15 194 111
4 0 167 16 194 167
5 0 222 17 194 222
6 0 278 18 194 278
7 083 0 19 305 0
8 083 056 20 305 056
9 083 111 21 305 111
10 083 167 22 305 167
11 083 222 23 305 222
12 0 83 278 24 305 278
Per cent of 18 g milk
Represents approximate BDI levels of 0832 Triton X-100 and 194 Na Tetraphosphate of 18 g milk
J
0~ 083middot
194 305 Figure 1 Variation in Sodium Tetraphosphate
21
Stability 2pound Reagent upon Storage and ~Relation
l2 Accuracy Experimental evidence concerning reagent -
stability at room temperature was provided by the previous
study on homogenized milk Table V provides a list of
reagent ages If the interpretation of the data is
correct one may assume that those reagents which provided
equal or significantly higher fat values than Mojonnier
ether extraction ie these reagents 3 17 25 and 58
days old could be considered equally reactive whereas
the detergent reagent which had been held at 350 C for two
weels in addition to being 64 days old and the excessively
old reagent (164 days) were not satisfactorily reactive
On the basis of these observations it would appear the
reagent is supoundficiently stable for test purposes up to at
least 2S days and probably so up to a month
Rate ~ Completeness gpound poundamp Liberation without
Centrifuging Sanders and Sager (19 1 p6) described but
did not recommend a detergent test which omitted centrishy
fuging In this study an attempt was made to determine
how long it would be necessary to hold a sample in the
tempering bath to allow for all the rat to reach the
graduated column To accomplish this two samples of raw
whole milk were prepared and reacted with the BDI reagent~
Following de-emulsification 50 per cent methanol was
added to bring the rat to the r eglon immediately below the
22
TABLE V
STABILITY OF REAGENTS
BDI Re411gant No Preparation Treatment
1 From 20~ Triton ~-100 Age at room and 22 Na Tetra temperature
2 17 Original materials
3 25 From 10 Na Tetra and 15 T X-100
4 58 Otoig1nal raateri_als
5 64 Original matariala Held 14 of 64 days at 35deg c
5 126 Original materials Age at room temperature
graduated column or the bottle Distilled water was then
added to raise the fat into the graduated column arter
which the s middotamples were placed in a tempering bath at soo C
Each of the tests was then read simultaneously every two
to three minutes until all the fat had risen Tvio other
normally centriiuRed detergent tests were used as standards
for comparison Table VI shOfS the progJgtess of the rat
gravitation Statistical analysis was not attempted
because the gravitation was complete even before the time
normally required for tempering the tests The lower
meniscus had not assumed its normal shapmiddote until after
about ten minutes Still these changes in shape did not
23
aCfeot the fat readings The original high r$a-d1ngs are
explained by the inclusion oi water b-etween fat globules
As the fat globulas collapsed the w$ter was expelled this
reduced the overall volume between the upper and lower
meniscuses
TABLE VI
OBSERVED RATE OF FAT RISE WITHOUT CENTRIFUGDlG ~LE
Time 1n ~ Fat Minutes l 2
0 4 so~lmiddot 400
3 435 middot4 54
5 4 38 4~40
7 4 42 440
9 4_ bull 40 4 30
11 4~ 40 4 40
13 4 bull 58 4 36
15 4 38 439
17 4 38 4 37
Centrifuged 4 38 4 38
Included ~ater between fat g lobules
(
24
J
GENERAL PHYSICAL AND CHEMICAL PHENOME1lA CONCERNING DE-EMULSIFICATION
Evaluation gt Detmiddotergent In this study an attempt
was made to find other nonionie detergents which might be
used successfully in the eliDllsion breaking tecbniqumiddote To
accomplish this the salt titration technique of Sager
18 pp 32middot38) was followed Briefly it involved titrashy
tion of a 10 per cent detergent solution with a saturated
sodium chloride solution at 55deg to 60deg C--the salt titexa
being obtained at the cloud po-int Following this proshy
cedure several nonionic detergent tolutions were
titrated The purpose or the experiment was to determine
if it were possible to so reduce the hydrophilie attraction
of the detergent by heating and salting that the lyophilic
attraction would predominate during the reaction middotperiod
It was observed by Sager that although some deterg ents
exhibit a cloud point with a minimum of salt and heat
they float on the surface This effect is obviously unshy
desirable since a detergent 11th sueh a specific gravity
wil~ tend to accompany the fat
only one type of compound of the group of detergents
studied (in addition to Triton X-100) demonstrated the
desired characteristics and was selected for further study
It was one of a series of Pluronics whose properties wmiddotill
be deseribed in further detail latermiddot
25
Studies ~ Washed Cream Preliminary investishy
gation with the first detergent method of the Bureau or
Dairy Industry which is based on the salt effect produced
somewhat erratic results It was observed that in some
instances sodium chloride used as recommended was no more
effective in promoting fat liberation than the detergent
alone Studies therefore were made ith ashed cream
in order to deter mine the reaction of the detergent and
salt with the rat globule in the absence of other milk
constituents A series of test tubes were prepared with
varying levels of sodium chloride added to Triton X-100
and to Pluronic L--64 in small volumes of was1ed cream
standardized in water to 4 per cent fat The tubes were
placed in a boiling water bath for 15 minutes In the case
of the Triton-salt mixtures it was observed that the detershy
gent was not completely dependent upon the presence of salt
for fat liberation although a small quantity was found to
be beneficial The Pluronic detergent was even less
responsive to the presence of salt although it was not as
inherently efficient as Triton in de-emulsification
Surface Tension Studies The role of surface tension
in the detergent test was considered from two different
standpoints The first stemmed from the early observation
that the shape and dimension of both the lower and the
upper meniscus of the fat column in the detergent test
26
differed from those of the Babcock This apparent difshy
ference is shown in Figure 2 The photographs are of a
test on the middots e mi lk sample in similar test bottles using
the two dipoundferent methods It was anticipated that
measurements of surface and interfacial tension ould
justify an explanation of themiddot phenomenon on the basis of
differences in spreading coefficients due to reduced
surface energy in t he case of the detergent test
The other consideration centered around the
theoretical aspects of the reaction mechanism by which the
fat is liberated The implications of the effect of a
surface- active agent upon the fat are obvious but there
has been no experimental substantiation of existing
theories concerning the actual detergent- fat interaction
It as with these factors in mind that the following
limited observations were made
Experimentation was based upon surface tension
measurements made with the du Nouy Interfacial Tensiometer
Readings were made at approximately 35deg C
The milk fats were prepared by reacting sulfuric
acid (Babcock) or Triton X~lOO (BDI method) with about 20
grams of 37 per cent cream in 18 per cent cream- test
bottles The two Babcock fat preparations were developed
by (a) the standard procedure plus 5 ml water to reduce
burning and (b) replacement of approximately one-half of
the underlying acid layer with distilled water so added as
27
Milk A Mil k A Milk B Detergent Babcock Detergent
Figure 2 A Comparison of the Meniscuses of Detergent and Babcock Fat Test Columns
28
to partially wash the extracted fat Detergent fat
samples were obtained by the following modified methods
a) The standard BDI procedure wit hout using methyl
alcohol or water (b) the standard procedure t he fat
being raised to the surface with water only (c) the
standard procedure the fat being raised into the neck of
t he bottle with methyl alcohol (50 per cent) and (d) the
standard procedure~ t h e fat being raised first with
methanol one-half of the lower layer then being syphoned
off and finally the rat being washed (or extracted) with
more methanol
The fats prepared as indicated and the underlying
liquids from the respective preparations were used to
obtain the surface and interfacial tension measurements
listed in Table VII The surface tension reduction
produced by the detergent is slight if it occurs at all
There is little difference either i n t he surface tensions
of the underlying liquids or in the interfacial tensions
between the aqueous and f at phases
Viscosity Changes in ~ Milk-Detergent Atlxtures It
was observed thst changi ng t he normal concentration of
sodium tetraphosphate in t he BDI r eagent altered the apshy
pearance of t he milk solids in t he body of the Ba bcock
bottle after heating This variation as not always
consistent even wi t h the same concentration of tetraphosphate
29
TABLE VII
SURFACE TENSI ON MEASURE~lENTS ON BABCOCK AND DETERGENT SYSTEMS 370 C
du Nouy Surface
Intershyfacial
Tension Tension Samshylli Treatment Fat4
Underlying Liq dypefem
FatLiq~4Z dyPe(cm
1 Babcock Reg Fat 3600 3700 428
2 ditto washed 35 82 35-90
3 middotBDI orlg fat (unextraeted) 5592 36 8
4 middot ditto + HgO raise rat
to 36 12
5 ditto + MeOH (50$ ) 3590
6 ditto bullwashed with MaOH 3610
7 middot BDI Reagent 3~ Triton X-100 7 Ns Tetra-phosphate) 36 bull40-ilshy
8 Cream 37 25~ Fat 46S~L~
9 Crealll bull equalvol BDI Reagent 402-iishy(7) (Not heated)
one reading -l$-gtJ-Avg of 5 readings
(Avg of shy3 readings
30
but tJU alao d~pondent upon t-~ solids-nottat content of
the ntlk fhera w however a trend in the ppearanea
ot middotthe milk rQaidubullbull thamp moSt obvou $station boog
that or gelatirloua form-ation ( thixotro-p1c jel vitb the( milk solids at rn1nism1m tetraphosphnto coneentrations
~ obviou 11npo-Jtt of suoh a -viampeou$ f-ortlltt on would
be the pr W tion opound SOU$ of the f t obul(JS from roachshy
ing tho gladuatbulld column cpound the bottlamp the-reby ~u_aing low
retuiingz ima phenomtJnon ia not tmltke tbe formetion o-f
viscous complexets- ot protem with anionic d tergGnta or
aetuill complsx r~t1on at pbospho~ic a-ed and pcentl7bull
phoaphste salt-s with protein
A dirampot appJroach waamp ~ to t middothamp cons1der-at1 on tgtf
vtseosity chlnga$ bf mea uring tm viscosit7 or- the
telc-t1on mtttweamp prior to lllld after ba-ating w n the
(Hlneentrat1on of sodium tetrapbospbate was varied Readshy
ings were rnareg at 60deg e- in watel bath oontrtltJUod at
1 c s1nee this is th~ tinel equil1b~um tempairu~ tti~
~aaing the test Two Ostwald vtscosmeta~s middot a~ atQildardbull
1zad with tr1plamp-d1stilled water at soO cjlj Spaoit-1c
Vity mafuuN~s ~la obtained by means ot a Wo$tpbal
balanee tuung tripledfJtilled water as raterance oaeh
titlamp moasur ents were de In addition to tho metlamp~
nmnts made on mUkbull samples of supercentrifuged skim milk
(20000 rpm) wet$ alao- er _ad The obaervati ans made on
tho methanOl phase were donG itlt ext~emamp eve t-o pre-11ampnt
Sodium tetraphn~phate per cent Figure 3 Influence of Na-tetraphosphate on Viscosity
32
mixing with the milk phase although the more fluid samples
at extremes of tetraphosphate concentration were diificult
to contr-ol
As may be observed in Table VIII and Figure 3 there
is a slight variation in the viscosity of different samples
of milk at the same sodium tetraphosphate level although
the shape of the eurve remains approximately the same
Starting from zero level tbere is an original minimal
value folloved by a sharp rise which exoaeded the capacity
of the viscosimeters available The extreme in viscosity
is followed by a symmetrical decline followed by a straight
line constant relationship with further increments in
poJyphosphate
Suggestion that a test for fat could be developed
without the use of the centrifuge raisod the question of
the iDrluance of the viseos1ty of the 50 per cent methyl
alcohol solution which is used to float the fat If careshy
fully layer-ed over the reaction mixture the methyl
alcohol can be observed in a very clear distinct phase
with very little intermixing of the milk constituents It
had been noted that the degree of clarity of the methanol
phase and the ease with which it could be attained in difshy
ferent tests depended again on the concentration of tetra-oi-J
phosphate and also the relative solidsmiddotnot- iat of the 1
particular milk Those tests in which a minimal quantity
of tetraphosphate bad been used (0 55) invariably had the
v
TABLE VIII
VISCOSITY DETERMI NATIOtgtTS ON VARIOUS DETERGENT- MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Tim$ Absolute Vis cos at Tetra- Triton meter in Vis cos Relative
SsJnple 600 c Qh OS ~ X-100 ~ I or II ~ c Ebull to H2Q
l Triple dist s2o 0~9938 0 0 I 511 4699 100
~2 n It o 9938 0 0 II 776 4699 100 3 50 Methyl A1o 0_9008 0 0 I 824 6 94 148
4 dup r 834 702 149 5 Skim milk 10137 0 0 133t 699 6 27
116 (l8ml) bull BDI (5ml) 10202 194 oss I 895 8 54 182
7 (lS) BDI (5) MeOH (25) 09638 194 oes II 13416 s 2a 1 76
a SUper Skim 10124 0 0 II 10347 sss 135
9 (18) BDI (5) 10192 194 oaz II 12221 7 52 160
10 fl middot(18) fl (5) MeOH (25) Ot963S 194 083 I 7990 747 159
BDI Reagent 7t$ Na Tetraphosphate 3 Triton X-100 (A middotUI
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- Time Absolute Viscos at Tetra- Triton meter in Viscos Relative
Sample 60deg c phos Xo-100 I or li sec c Pbull to H20-11 Milk (A) 10088 0 0 II 1093 679 144
12 10125 0 083 I 748 700 149 13 10201 194 0 I 1007 960 204 14 10223 194 0~83 II 1442 907 193
15 - 056 oas II Too viscous to measure 16 n 10192 139 oas I 1159 1104 235
17 10250 250 083 II 1466 925 197 18 ff 10265 3oa 083 I 94 9 911 194
19 Milk (B) 1013 028 083 II 2761 1722 366
20 083 oa3 --- --shy-lHtSample too viscous to measure with Westphal balance or Ostwald viscosimeter
(Jl ~
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIONS ON VARIOUS DETERGENT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Visoosibull Time Absolute Viscos a~ Tetra Triton metexo in Viscos Relative
Sample 60 c phos X-100 2f I or Il see 0
to HaOPbull- bull
21 Milk (B) 1913 l~tll 083 bullbull middot-middot 22 n 1 P22 167 ft I 967 924 1971
gt
23 lP26 2~2 II 1378 870 185
24 1030 2 78 I 911 877 187 bullbull 25 ff 1054 333 II 1394 887 189
26 1058 sae rt I 920 8 93 190 27 n (B) bull BDI (5)
umeOH (25) 09820 1 94 II 1387 see 1139
28 21 MeOH (25) 09416 111 II I 1258 1)063 255
29 24 MeOH (25) 09509 533 Jt I 923 887 189
~kSpeeific gravity extrapolated from straight-line relationship between specificgravity and per cent No tetraphosphate added to those samples whieh could be measured
TABLE VIII (CONTINUED)
VISCOSITY DETERMINATIOtlS ON VARIOUS DETERGEliT-MILK SYSTEMS
Ostwald Mean Sp Gr Na Viscosi- 1ime Abaolute Via-cos at Tetrabull Triton meter in Viscos Relative
Sample sect0~0-~ P~bo~_ $ X-100 I or II ~ C Pbull - ~2~~2()__
30 Methanol phase over Milk (C) 09310~028 0-83 I 912 middot794 1 bull69
middotshy31 n 09027 056 u _I 1308 727 1 bull 55 32
~
09002 083 I 855 719 153 - 33 tf 09091 097 II 1313 734 156
34 0 bull 9164 167 It I 8843middot 5 161 bull
11 35 09167 194 II 13243 7 47 1_59 bull
36 Silk C without Methanol 10125 oza tt II 1417 958 204
37 Milk c (18)
Methanol ( 25) 1_0132 0 0 I 832 788 168
Higher spacitle gravity aasqc1ated with intermixing ofmilk solids with alcohol
middot(If m
37
clearest methanol layer and very seldom was any unreacted
cream observed in the fat column
It was not surprising therefore to find viscosity
measurements of this methanol phase following a pattern
almost inversely that of the same reaction mixture but to
a less exaggerated extent Actually the observed differshy
ences in viscosity of this phase at various levels of
tetraphosphate would seem to be so slight as to have little
effect upon the rate of fat rise with or without centrishy
fuging
Figures 4 and 5 show the typical precipitate formation
and changes in protein hydration which accompanies variashy
tion in sodium tetraphosphate concentrations
Nature 2pound ~ Tetraphosphate Precip-itate Several
different approaches were made to the study of the nature
of the complex which was observed t o be formed when the
concentration of sodium tetraphosphate in the milk exceeded
about 09 per cent Vbile the observations presented offer
some plausible explanation the investigation was not sufshy
ficiently extensive to furnish definite proof of either the
composition or the precipitate or the mode of formation
However since it was subsequently found that precipitate
formation had little bearing on the quantitative aspects of
the test further work did not seem justified
It was firmiddotst thought that the precipitate illustrated
38
by Figures 41 5 and 6 might be largely proteinaceous and
ot the type involving the serum p~oteins This concept
was suggested by the numerous observations which confirm
the formation of irreversible complexes by anionic detershy
gents and polyphosphates with albumin 1 This complex has
been observed to form a precipitate at pH values acid to
the isoeleotric point where the precipitant was not
present in excess ie where the detergent-protein ratio
did not exceed 075 Since the pH for most of the reaction
mixtures was not acid to the isoelectric point of any of
the milk proteins (most of the pH values for the reaction
mixtures were between 69 and 74 with a mean of 725)
this suggestion has little theoretical support Furthershy
more the fo l lowing experiments demonstrated that little
if any protein was involved in the precipitate formation
An estimate of the protein distribution in a milk
sample following reaction with the tetraphosphate was
obtained by determining the Kjeldahl nitrogen on the upper
most 5 ml of t he samples studied in the viscosity experishy
ment The observed values are listed in Table IX Any
differences in the protein values of these samples with
variable sodium tetraphosphate are believed to be within
the limits of experimental error
lsee Biblioe5raphy (1 pp256-257)(3 ppl0-12)(7 pp 2854-2855)8 ppl83-185)(9 p~317-324)(10 pp483-486)(12 pp400-407)(13 pp l6-20)(15 pp709-710)17 pp l95shy209
010 015 020 025 030 035 040 0 45 050 Sodium Tetraphosphate g per 18 g milk
Figure 4 The Effect of Variation of Sodium Tetraphosphate upon Protein Hydration in
Detergent Tests
Percent 028 083 111 167 222 278 333 Natet
Figure 5 Inorganic Precipitate Formation with Excesses of Sodium Tetraphosphate in Milk
41
Figure 6 Compari~on of Detergent Tests in which 50 Methanol (1 ft) and Distilled Water (right)
were used to FlQa~ the Fat into the Neck of the Bottle
42
TABLE IX
RELATIONSHIP BETWEEN PROTEIN DISTRIBUTION AND PRECIPITATE FORMATION WITH VARIATION
IN POL[PHOSPHATE CONCENTRATION
Source of Na Tetra gt SamEle SamEle EhOSEbate ~ Protein
1 Milk 5 ml top portion 028 264 after heating in H20 bath
2 o83 261
3 111 259 4 167 256 5 278 257
6 389 246 7 rihey (097 028 078 a total protein) If 056 076
9 n tt 111 078
10 167 No value tt rJ11 222 078
12 278 No value 13 333 079 14 Milk lop Sml 019 294
15 It Middle 5 ml 019 289
u16 Bottom 019 307
17 Composite 019 284 18 Mixed Whol e 0 19 289
43
A sample of cheddar cheese whey which had developed
about 0 4 per cent acidity calculated as lactic was
treated with sodium tetraphosphate the concentrations
varying from 0 to 06 g18 g whey in 200 mm test tubes
which were then heated 15 minutes in a boiling water bath
No precipitate was observed If the complex formation
were one involving serum proteins solely it vould seem
that a precipitate should surely develop at the pH range
observed (538 to 6 42) Determinations of protein
n1trogen on the uppermost regions of these preparations
showed the protein distribution to be the same as the
preceding experiment These values are also listed in
Table IX
Sediment from supercentrifuged milk (42000 rpm) was
redispersed to give a suspension containing 4 per cent toshy
tal solids which lere assumed to be composed largely Omiddotf
calcium caseinate and calcium phosphate salts Applying
the same technique as before it was observed that again no
precipitate formed The pH of the redispersed sediment was
69 and the same observations were made on a sample of the
same material the pH of which was adjusted to 6 52 Howshy
ever when methyl alcohol was added and the sample rehented
a precipitate was formed This latter observation did not
suggest any solution to the problem since the normal
precipitate ormation was observed in whole milk even
before the methanol had been added It did appear that the
44
methanol might contribute to such formation in the typical
test although the temperature after reaetion does not
exceed 60deg c More conclusive evidence was obtained with the
preparation of 134 g milk and 2 59 g sodium tetraphosphate
which were reacted in a 2 x 12 in test tube The
precipitate was allowed to settle for 45 minutes after
which protein determinations were made on the top middle
and bottom portions as well as on a composite of the
whole From the data in Table IX it is seen that the
precipitate accounts for only a small portion of the total
protein
The final attempt to separate and analyze the
precipitate involved the preparation of 180 g of milk
6 g of sodium tetraphosphate 3 g of Triton X- 100 and
water to make a total opound 237 g The resulting precipitate
was allowed to separate the supernatant liquid was
decanted and finally the precipitate was centrifuged out
of suspension at 2 000 rpm for 15 minutes The precipitate
was analyzed for total solids protein casein and ash
Results of this preliminary analysis indicate the
precipitate contained roughly 77 per cent total solids 13
per cent total protein (moisture-free basis) 10-12 per
cent casein on same basis and 48- 52 per cent ash The
reaction mixture prior to heating contained 1416 per cent
total solids 3 per cent ash and 3 91 per cent protein or
45
28 per cent protein on a moisture-free basis Two milks
from different sources yielded centrifuged precipitates
weighing 5 25 g and 842 g respectively prior to drying
Since sodium tetraphosphate is highly soluble it is doubtshy
ful if any of this precipitate is pure insoluble tetrashy
phosphate
From this series of experirents it appears that the
precipitate found occurring at higher concentration of
sodium tetraphosphate for the most part is an inorganic
precipitate presumably of calcium salts and does not inshy
volve insoluble complex formation with protein
Action 2 Methyl Alcohol The action of methyl
alcohol in the detergent test appears to be dependent upon
its ability to free the fat of absorbed detergent or other
surface-active material which tends to dissipate such
surface energy as serves to maintain the fat on themiddot
surface in a dispersed form When the 50 per cent methanol
solution is added after the de- emulsification of the fat
the smaller g lobules of dispersed fat seek to combine into
larger masses with a resultant conservation of surface
energy In this manner milk solids or other material
adsorbed to the numerous small fat globules are kept from
being swept into the fat column Furthermore in instances
where methanol is not used there is often a small tuzz
of unreacted cream on the bottom of the fat column after
46
centrifuging This fuzz is greatly reduced and often
eliminated by use of the methanol The comparison of
tests developed with methanol and water is shown in
Figure 6
The capacity of the methanol to clarify the fat
c olumn was applied in modification of the original BDI
procedure to further improve the efficiency of deshy
emulsification It was found that the addition of small
amount of 50 per cent methanol during the later stages of
the debullennllsification process materially increased the
clarity of the resultant fat column and greatly insured
against cream globules escaping de-emulsification The
technique used was to squirt about 1 ml of the methanol
solution from a wash bottle into the test bottle five
minutes prior to the end of the heating period The
separate fat globules would immediately spring together as
the fat collected into one mass After doing this the
bottle was given a slight swirl if any unreacted cream was
observed on the surface
~ 2[ ~ ~ Observe Progress of De-emulsification
In the course of performing the detergent analyses it was
found that a small amount of dye (methlyene blue) added to
the BDI reagent aided in the accounting of those samples
to which the detergent had been added During the heating
period it was observed that the blue color was adsorbed
47
on the unreacted cream globules at the surface of the
reaction mixture As the de-enntlsiiicat1on reached comshy
pletion the dye moved i nto the aqueous phase leaving the
clear fat free of dye It was observed that any test in
which the dye remained~ in the interface required either
additional heating o~ the fat needed to be re-mixed with
the detergent for further reaction If color was observed
at the fat - water interface and the test were treated as
having reached completion there was usually observed a
cream ufuzz at the bottom of the fat column after centrishy
fuging
This phenomenon which occurred so consistently with
the first two batches of reagent could not be faithfully
repeated in later i nstances It is believed that some
actors concerned with the aging of the dye in the prepared
reagent could have been responsible for the effect It
would certainly be an asset to the test if those mechanisms
could be explained
Nevertheless whether the dye performs in the manner
indicated above or not the colored background provided
for the light colored cream globules is beneficial enough
to justify the use of the dye Figure 7 shows the effect
or dye adsorbed to the cream
rshy
I
I
Dye adsorbed on surface cream After 15 minutes dye is demiddot during early stage or hoating adsorbed emulsion is broken
Figure 7 The Adsorption and De-adsorption of Methylene-blue Dye from the Surface of the Detorgent
Teat During Fat Liberation ~ (X)
49
Studies Involving Pluronic Burfa~e-Act1ve Agents
The Pluronic series of surface-active agent$ presented an
interesting group of materials for study of the demiddot
emulsification process These compounds are formed rrom
the condensation of ethylene oxide with the product
resulting from the previous condensation of propylene with
propylene glycol This gives the aeries an emperioal
rormula of HO (OgH4)a(C3H00)b(02H4lcHbull In the three comshy
pounds studied-middotPluronic L-64 L-62 and F-68--the ethylene
oxide content increases from 20 to 80 per cent The preshy
fix L denotes liquid and F represents the powdered
state whimiddotoh is due to this compound containing the highest
proportion of the ethylene oxide molecules The nunieral
abull denotes that these compounds have an average molecular
weight of 600 Thus it can be seen that these compounds
can be tailor-made to produce a surface-active agent
with the desired hydroph1lle l1poph111c balance
These surface-aotive agents responded favorably t o
the salt titration demonstrating a low fat-solubility
and the capacity of the L-64 and L-62 members to be salted
out or solution upon the application of heat In the debull
emulsification or milk fat L-64 was found to possess the
most desirable hydrophilic-lipophilic relationships
Further experimentation brought forth several observations
which suggested how these groups of materials_might be
substituted for Triton X-100
50
When these three agents were used separately only
L-64 was found to effectively de-emulsify milk Even so
the reaction was not complete when this material was used
with salt according to the first technique of Sager
ashed cream was used in the case of Triton X-100 in order
to determine the action of the detergent without the
interference of proteins or other milk constituents It
was discovered that Pluronie L-64 would liberate all but a
small quantity of the fat from the cream to a large
extent independmiddotently of the presence of salt Combinashy
tions of L-64 with L-62 gave no improvement but when a
small quantity of Pluronic Fbull68 was added to the L-64 the
last remaining trace of unreacted cream was de-emulsified
thereby producing a clear column From these data a
combined reagent of 8 per cent Ir-64 5 par cent F-68 and
3 per cent sodium chloride was formulated w~ ch when apshy
plied to the testing of whole milk by the usual procedure
gave clear tests with a good correlation with Babcock
values Application of the combined reagent to homoshy
genized milk was not as successful even after prolonged
heating the resulting fat columns contained much unreacted
cream Attempts to repeat the results obtained with whole
milk were not entirely satisfactory
In the course of these experiments several technishy
ques were used in attempt to i ncrease the efficiency of
51
the fat liberation Some of these procedures may find apshy
plication in future test methods but in this work they
were not extensively studied and therefore represent only
suggestions In this connection it was found that
diluting the milk with 10 to 15 m1 water before heating
seemed to improve the fat liberation Ethyl alcohol was
more efficient in clarifying than methyl when added to the
test bottle before the end of the heating period It was
observed that when the methyl alcohol solution was added
in two installment~ each followed by centrifuging
invariably the fat collected upon the sloping sides of the
test bottle and would not be forded in the graduated area
of the test bottle It was found that test bottles had to
be thoroughly cleaned to prevent the fat from adhering to
the sides of the bottle Thorough cleaning is equally
applicable to do-emulsification using Triton X-100
52
DISCUSSION
Statistical Analyses The accuracy opound the Bureau opound
Dairy Industry detergent test was found to be equivalent
to that of the official Babcock when testing raw whole
milks of varying fat contents The extreme range in fat
content was provided for by the use of the Jersey and
Ho~stein breeds for the samples On the basis of this
information one may conclude that the properly performed
detergent test liberates all the fat of normal whole milk
with no measurable solution of the detergent in the fat
With homogenized whole milk the detergent method
yields values which have a positive deviation from the
Mojonnier ether-extraction method similar to the normal
relationship e~sting between official Babcock and Roeseshy
Gottlieb with whole milk Although further studies are
indicated in order to fully establish the relationship
between the quantitative fat liberation by detergent and
the ether-extraction principle one may conclude that the
detergent method is fully as efficient as the modified
Babcock (acid added in three installments in testing
homogenized milk Moreover it appears from the study that
the detergent method is more efficient than the modified
Babcock
A fairly wide range of detergent concentrations was
found to be effective for de-emulsifying milk of medium
53
fat content The sodium tetraphosphate was found to be
non-ossential for detergent action although its presence
was round to be desirable for dispersing the proteins and
salts and preventing protein coagulation during heating
By removing the milk solids from the alcoholic solution
below the fat its use enabled a more accurate definition
of the lower meniscus
Neither reducing the quantity of Triton ~100 below
the BDI recommended level of 015 to 010 g per 18 g milk
nor increasing the detergent content beyond 0 5 g per 18 g
milk had any deleterious effect upon detergent action
This means that the detergent is equally effective over a
range of 056 to 278 per cent of the 18 g milk This is
significant from the standpoint of reagent preparation in
that there vtould be an allowable variation of from 062
to 32 fluid ounces Triton X-100 per quart where an optishy
mum of one fluid ounce was desired
The reagent has good stability in storage at normal
room temperature Any deterioration seems to involve the
sodium tetraphosphate to the greatest extent Vfllen such
exhausted reagent is used in a test there occurs a
decrease in pH and an increase in the hydration of proteins
similar to the effect of low Tetraphos 11 concentration The
activity of the Triton appears little affected by these
changes in the polyphosphate It might be possible to
renew the activity of any reagent older than one month by
54
the addition of a small quantitv of fresh sodium tetrashy
phosphate Under normal circumstances fleeh reagent
should be prepared monthly
The success with which the detergent method can be
applied t -o a fat test where centrifuging is not practiced
is mainly dependent upon complete de-emulsification of the
cream and the preventim1 of fat entrapment in the viscous
mass of milk protein resulting fr-om tetraphosphate
deterioration or insufficiency In this type of test
extra attention should be given to the cleanliness of the
test bottles sincemiddot there is a marked tendency for th-e fat
to cling to the sloping sides of the bottle The tests of
this type which were studied indicate that an accurate
reading can be made within seven minutes after the
methanol and distilled middotwater have been added to raise the
fat into the graduated column~ although the shape of the
~ower meniscus does not assume its final form until about
ten minutes have elapsed Since a statistical analysis
of this type of test Vas not attempted the absolute acshy
curacy of the method is not established It appears aae
to state however that if the detergent has completely
da-emulsif1ed all the rat and sodium tetraphosphate has
completely dispersed the milk proteins the test can be
read with confidenc~ after 15 minutes in the tempering
bath
55
General Phenomenon Related 12 Detergent Action
Althougl-1 the experiments were not designed specifically
to determine middotthe reaction meellSnism by which the tat
emulsion is broken the accumulated obselvations provide
some basis for theorizing on the interaction between the
detergent and the fat
The salt titration middotwas used by Sagsr to establi~h
fundamental criteria for those detergents used to deshy
emulsiiy milk fat To meet these requirements it is
necessary that the detergent possess a hydrophilicshy
hydrophobic balance that upon heating its strong hydro
phtlic attraction can be reduced with the result that it
will be more strongly attracte-d to the fat Sager inshy
ferred that a salt further suppresses the hydrophilic
pro-perties of the deter gent and thus aids in fotming an
undefined complex with the fat which then rises to the
surface Upon dilution Pfith methanol this complex disshy
sociates thus allowing the liberated fat tomiddot coalesce
The effect of salt upon the activity of Triton X-100
was studied with washed cream which was a system from
whieh tho osmotic or- salt effects of other milk constitshy
uents had been removed It was observed that salt aided
but did not completely control the de-emulsification
activity of the detergent in this otherwise salt-free
medium- With milk however the effect of salt is apshy
parently supplanted by the natural salts of milk and as a
56
result neither salt nor sodium tetraphosphate is
required to effect de-emulsirication
As theorized by Schain (20 pl22) the breaking of
the milk emulsion is dependent upon the formatfon ot a
complex between the protein of the fat globule membrane
and the anionic detergent This concept does little to
explain the emulsion-breaking capaei t y of nonionic detershy
gents which do not appear capable of forming complexes
with protein Furthermore if protein-complex formation
were the key~ one would expect that sodium tetraphosphate
which combines stoichiometrically with protein in effectshy
ing its dispersion would be capable of de-emulsification
This is not the case Numerous other anionic detergents
have ahlost no capacity for de-emulsification even though
they are capable of reacting with protein on an ionic
basis
One cannot ignore the effectiveness of the numerous
anionic detergents reported by Patton and Stein (22) to
be capable of de-emulsifying cream when added in suitable
solvents It would appear that although such factors as
protein dispersion and membrane protein combination with
anionic materials are an aid to de- emulsification by
detergents they are not the prime factors
Since disruption of the fat globule membrane through
complex formation with membrane protein does not appear to
57
be a satisfactopY explanation of de--emulsification other
mechanisms should be considered One of these is based
on the theory of Sager and ca-workers that salt (sodium
chloride or sodium tetraphospbate) and heating cause a
shiftmiddot in hydrophilic balance or the detePgent middotA complex
between the fat and the detexrgmiddotent is produced by the heat
ing and later is broken by the cooling and the dilution
with the 50 per cent methanol During the reaction
period the fat-detergent complex floats to the surface
In accordance with the latter theory it would
appear that duxoing this complex formation the detergent
replacement of the natural milk emulsifier ( phospholipidbull
protein complex) which originally served to diasipate the
high surface energy of the fat g lobule It would seem
essential that this material be rep~aeed sinee neither
heating at 100deg C nor the action opound sodium tetlaphosphate
is sufficient to break the emulsion although the protein
associated with the phospholipid is likely denatured in
the process
One of the most logical explanations of this
removal of membrane material is based on the surface
tension reduction capacity of the detergent If dUring
the heating the detergent --JUan reduce the surfa~e energy
ot the fat globule interface to a greater extant than
does the natural membrane it will be prbullefe~entially
adsorbed The Triton X-100 which is not an efficient
58
emulsifier however due to its own normally low affinity
for fat leaves the fat globule upon cooling and dilution
(Sager As a result the fat globules are left with a
high surface energy which is dissipated by coalescence
It would be desirable to obtain surface tension data
at the fat globule interface to shed more light on the
actual energy ehange which may lead to the de-emulsifieashy
tion Unfortunately the data obtained in this work only
apply to a comparison of detergent extracted fat with
Babcock and extracted fat and give no information of the
state of the natural fat globule prior to and during the
actual fat release
The extreme increase in vis-cosity of the reaction
mixture which accompanies a decrease 1n the sodium tetrabull
phosphate concentration to a level of approximately 0 6
per cent can be a source of error in the detergent
analysis This was best demonstrated by the effeet upon
the values obtained in the experiment concerned with
stability of the reagent In thia experiment it was found
that the excessively old reagents acted similarly to reshy
agents containing a minimal quantity of tetraphosphate
Both were responsible for the formation of a viscous mass
of hydrated protein in the body of the bottle As an
indication of the cohesion of this mass the methyl alcohol
layer was observed to be almost perfectly clear as was
also the fat which escaped from the viscous mass It
59
becomes obvious therefore why these tests were obsmiddoterved
to hav-e lower taadings
This fat entrainment
was not observed in the experishy
ment on the affect of variation of reagent concentration
in which centrifuging was employed This was probably
due to the fact that the lowest level of tetraphosphate
used was 083 per oent and the viseosity of the mass was
not sufficient to overcome the extra gravitational effect
of the centrifuge These observations do not justify
the recommendation that low tetraphosphate eoncentrations
be ignored- 1he viscosity data provide evermiddoty indication
that this will invite incomplete libex-ation
The formation of a white precipitate with excessive
tetraphosphate concentrations is of little significance
insofar as the quantitative character of the test is
concerned All indications suggest that the precipitate
forms as a result of the mass ion effect of polyphosphate
excess The very low pt-otein content of the precipitate
indicates that protein is not involvod in the reaction but
is mechanically occluded
It was observed that the detergent test tertds to be
read slightly higher than the standard Baboook This
suggests that the upper meniscus is more concave in the
formex- This difference was not great enough to be
detected by the statistical analyses when all the tests
were read by a single individual It appears reasonable
60
that the fat freed by the detergent test ould have a
greater tendency to spread or wet the glass of the test
bottle at the upper menis cus th~~ the fat in the Babcock
test That 1st the work of adhesion between the fat and
the neck or the bottle minus the work of the cohesion of
the fat would be a more positive value for the detergent
test and the spreading of fat on glass would be greater
The v~sual manifestation of this effect would be that the
contact anglo between the fat and the glass would be less
in the ease of the detergent test thereby accounting for
the more concave appearance of the detergent meniscus
~e few observations of the surface and interfacial
tensions of the B~bcock and detergent test fat failed to
show any significant differences which could account for
this phenomenon insofar as the energy relationships between
the fat and the glass wall were concerned Other factors
such as the possible presence or detergent on the alls of
the bottle may complicate the situation The same factors
may apply to the appearance of the lower meniscus In adshy
dition the fact that there is a difference in the specific
arav1ty of the aqueous phase Lmmediately below the fat of
the two tests may account for differences in shape of the
lower meniscus
The two techniques were found to improve the efshy
ficiency of the fat liberation One involvod adding a
small quantity of the methanol solution to the fat during
61
the later stages of the de-emulsification This increased
the clarity of the liberated iat and reduced the time
required for the reaction
The other modification makes use of the adsorption
of methylene blue dye on the unbroken cream emulsion to
determine the degree of de-emulsification as the heating
in the water bath progresses The factors governing the
dye adsorption middotstill remain to be determined if the
phenomenon is to be made consistent Nevertheless~ the
use ot the dye as a background for the unreaeted cream
globules is useful enough to justify its adoption It is
unlikely that its use would jeopardize the precision of
the detergent analysis
While the results with Pluronics were not entirely
satisfaetory from the standpoint of quantitativmiddote fat
liberation the action of these different compounds did
suggest p-ossible contbinations which might be more effective
than single compounds For example it was tcrund that
the addition of a small quantity or Pluron1c F-68 to the
Ir-64 agent produced a more eifmiddotecti v~ combination than the
use or L-64 singl-y It would appear that F-68 is too
hydrophilic in its makeup (containing 80 par middot cent ethylene
oxide) and L-64 is not sufficiently hydrophilic (40 per
cent ethylene oxide) bull The most desirable balance seell$
to be attained when the control of ethylene oxide is
a pproximately 50 per cent The proper relationship of
62
ethylene oxide to the lipophilic residue would have to be
estmiddotablished before these materials would have sufficient
reactivity to de--emulsify milk
63
SUM1MRY
The detergent method of the Bureau of Dairy Industry
(BDI) was found to be equivalent in accuracy to the
official Babcock procedure when testing raw whole milk
of normal fat content It gave significantly higher
results than the Roese-Gottlieb (Mojonnier) method in the
case of homogenized milk and still higher results than the
modified Babcock which gave significantly lower results
than Mojonnier
Triton ~100 was found to be effective over a wide
range of concentrations in de-emulsifying milk of medium
fat content Sodium tetraphosphate was found to be of no
significance in the BDI test from a quantitative standshy
point unless present in a low concentration Under the
latter conditions the tetraphosphate reacts with the
protein with the resultant formation of a highly viscous
mass which entraps fat globules and thus contributes to
low tests
The prepared reagent remains stable for at least six
weeks at room temperature after which the sodium tetrashy
phosphate appears to deteriorate The activity of the
Triton X-100 does not seem to be affected by this
deterioration The deteriorated reagent will de-emulsify
the fat but the viscosity of the hydrated protein prevents
some of the fat from rising to the surface This is the
64
same phenomenon which occurs in the teats with low tetrashy
phosphate concentrations when the BDI test is not centrishy
fuged
It appears that most of the rat rises within the
first few minutes after the methanol solution and water
have been added to float the fat An accurate tes t not
requiring centrifuging is possible providing extra
attention is given to the cleanliness of the glassware and
to completeness of reaction with the detergent Centrishy
fuging however is to be recommended
Both the upper and the lower fat meniscus of the
detergent test appear to be more concave and convex
respectively than the same in the Babcock The surrace
and interfacial tensions for the detergent-fat system
appear to be the same as the Babcock and that-afore provide
no basis upon which the apparent differences in spreading
of the fat can be explained
The progress of de-emulsification can be more closeshy
ly followed by adding a water-soluble methylene-blue dye
to the reagent so as to provide a dark background for the
c~eam particles The addition of a small quantity of the
methanol to the test bottle before the end of the heating
period increases the clarity or the fat and aids in
perfection or the BDI test
Studies with the Pluronic surtace-a~tive agents
indicate that if the proper balance of the ethylene oxide
65shy
group and the lipophilic residua can be attained sueh
agents would be capable of quanti tat1ve de-emulsification
66
BIBLIOGRAPHY
1 Anson M L The denaturation of proteins by synthetic detergents and bile salts Science 90256-257 1939
2 Association of official agricultural chemists Official methods 7thed Washington DC 1950 910p
3 Bull Henry B Mixed monolayers of egg albumin and lauryl sulfate Journal American chemical society 6710-12 1945
4 Greshenfeld Louis and Bernar d J Ucko Fat detershymination in milk Journal milk and food technology 13175-176 1950
5 Greshenfeld Louis and Bernard J Ucko Fat detershymination in milk and milk products ~ Journal milk and food technology 13342-343 1950
6 Greshenfeld Louis and Marvin H Rosenthal Fat determination in milk and milk products Journal milk and food technology 1417-18 1951
7 Lundgreen Harold P Formation of fibers from nonshyfibr-ous native proteins Journal American chemical society 632854-2855 1941
8 Lundgreen Harold P Daniel W Elan and Richard A 0 1 connell Electrophoretic study of the action of alkyl bonzene s ulfonate detergents on egg albumin Journal biological chemiotry 149183shy193 1943
9 Lundg~een Harold P Synthetic fibers made from proteins In Advances in protein chemistry Vol 5 N Y N Y Acade~ic press 1949 pp 317-324
10 Miller Gail Lorenz and Kjell J I Anderson Ultrashycentripoundug e and diffusion studies on native and reduced insulin in dupanol solutions Journal of biological chemistry 144475-486 1944
67
BIBLIOGRAPHY (CONTINUED)
11 Mojonnier brothers company Instruction manual for setting up and operating the Mojonnier milk tester Chicago Ill 1925 7lp (ItsBulletin no lOlJ
12 Neurath Hans and Frank W Putnam Interaction beshytween proteins and synthetic detergents III ~olecular kinetic studies of serum albumin-shysodium dodeoyl sulfate mixtures Journal of biological chemistry 160397-407 1945
13 Palmer K J The structure of an egg albumin detershygent complex Journal of biological chemistry 4812-20 194-4
14 Patton~ Stuart Preparation of milk fat I A studyof some organic compounds as de-ellDllsifying agents Journal of dairy science 35324-329 1952
15 Pearlman Gertrude E Combination of proteins and metaphosphoric acid Journal of bio~ogical chemistry 137707-711 1941
16 Putnam Frank w and Hans Neurath The precipitationof protein by synthetic detergents Journal of the American ehem1eal society 66692-697 1944
17 Putnam Frank w and Hans Neurath Interaction beshytween proteins and synthetic detergents II Electrophoretic analysis of serum albumin and sodium dodecyl sulfate mixtures Journal biological chemistry 159195-209 1945
18 sager Oscar S Fred l4 Grant and Thomas M Hammond A study of the Schain butterfat test In proshyceedings of the 44th annual convention of the milk industries foundation 1951 Laboratorysection pp 27-45
19 Sager Oscar s and George P Sanders A BDI detershygent t est for butterfat in milk and other dairy products a proceedings of the 45th annual convention of the milk industries foundat1on 1952 Laboratory section pp 1-14
68
BIBLIOGRAPHY (CONTINUED)
20 Schain Philip The use or detergents for quantishytative fat determination Scienee 110121middot122 1949
21 Schain~ Philip Determining the butter-rat content of milk Milk dealer 40(3)29505456 1950
22 Schain Philip Single solution method for detershymining butterfat in milk In Proceedings of the 43rd annual convention of the milk industries foundation 1950 Laboratory section middot pp 12middot17
23 Stein Charles M nnd Stuart Patton Preparation of milk rat II A new method of manufacturingbutteroil Journal of dairy science 35655-660 1952
24 Tukey John W Comparing individual means in the analysis of variance Biometrics 599-114 1949
25 Wi ldasin H- L E o Anderson and D E Watts A modification of the Babcock test employing a quaternary ammonium compound Journal of dairy science 3687-92 1953
26 Wildasin H L and E 0 Anderson A modified test for homogenized milk using cationic detergents Storrs Connecticut 1952 lOp (Connecticutagricultural experiment station bulletin 28)
TABLE X
FAT TEST VALUES FOR COMPARISON OF BDI METHOD WITH BABCOCK ACCORDING TO BREED AND COWS WITHIN EACH BREED
Cow No l~
BDI 2
Jersey
3 1 Babcock
2 3 l BDI
2
Holstein
3 l Babcock
2 3
l 6 44 6 43 6 42 5 53 6 33 6 37 4 86 5 02 493 4 93 4 88 472
2 7 17 7 37 6 68 7 29 7 32 7 23 4 90 4aa 4 70 4 83 4 78 4 72
3 aao 8 68 8 30 8 55 s sa sso s oo 4 88 sos 4 72 4 88 4 90
4 5 88 6 12 5 85 590 s ss eoo 4 16 4 25 4 25 4 17 418 4 14
5 6 63 6 85 th55 6 62 6~68 672 3 36 355 3 45 3 46 5 47 3 38
6 5 17 5 23 5 13 4 91 5 02 5 12 3 54 353 3 58 363 3 58 4 16
7 4 08 3 92 3 98 4 03 4 05 4 06 4 28 4 32 4 46 425 4 18 4 37
8 7 43 e seP7 sa 7 44 742 742 3 94 4 02 4 06 4 12 4 00 405
9 6 58 6 73 6 65 6 53 643 6 57 355 352 3 58 3 57 350 354
10 7 44 7 47 7 42 7 22 7 28 723 4 32 4 38 4 30 4 28 432 4 18
ll 6 bull 35 6 3B 6 20 6 33 633 625 3 62 ~ 52 353 359 367 358
12 7 02 7 22 6 98 715 7 03 6 92 4 25 4 04 ge 4 19 4 03 4 14
~~middotcomputed value for missing observation ())Series of replications (Q
70
TABLE II
FAT TEST VALUES FOR ANALYSIS OF EFFECT OF VARYING LEVELS OF SODIUM TETRAPHOSPHAPE AND TRITON X-100
01~ FAT LIBERATION
Sample No 1 2 5
Replieation 4 5 6 7 a
l 0 0 ( 0 0 0 0 0 2 0 0 0 0 0 0 0 0 s middot463 4 68 480 478 4 85 4 68 4 63 4-55 4 4 ~68 4 17 495 4 75 4 77 4 77 470 460 5 4 62 4 77 4 70 470 4 57 473 4 68 470 6 4 bull 60 4 63 4 bull44 468 462 457 4 56 455
7 0 0 0 0 0 0 0 0 8 465 4 middot48 4 75 4 ~ 50 4 60 450 430 470 9 4 70 4 65 4 53 4 75 4 73 4 76 4 75 475
10 4 72 4 65 4 47 4 68 4 73 473 4 70 4 ~78 11 4 64 4 6 middot4 38 4 68 4 80 4 69 468 455 12 4 60 4 55 4 bull53 4 55 4 65 4 62 4 68 458
13middot 0 0 0 0 0 0 0 0 14 4-78 4 83 470 4~ 82 4 ao 470 4 ~73 4~72 15 4 80 4 78 4 75 4 85 4 88 4 90 4 76 4-65 16 4 70 472 4-62 4 68 477 4 82 4 78 4 ~ 75 ~7 473 4 72 4 70 4 82 4 73 4 82 4 75 4 62 18 463 4 bull70 4 43 480 4 77 4 80 4 62 4 68
19 0 0 0 0 0 0 0 0 20 4 75 4 68 4 67 4 87 478 4 80 4-80 4 70 21 4 i42 3 68 4 45 4 85 4 77 4 75 4 80 480 22 375 4 76 4 33 4 83 4 76 4 80 469 472 23 472 480 445 4 65 4 75 478 4 68 473 24 3 02 4 68 4 58 4 78 475 474 472 4 64
71
II
TABLE XII
TEST VALUES FOR COMPARISON OF BDI BABCOCK llifD ROESE-GOTTLIEB (MOJOliNIER) WITH HOMOGENI ZED ~ITLK
Replications ethod I II III IV v VI
BDI 1 4 28 418 4 23 4 22 4 17 4 28
2 424 4 24 middot4 25 4-27 4 28 4 25
n 3 4 20 4 16 4 15 4 27 4 27 4 22
4 4 04 4 19 4 03 413 416 4 16
ft 5 4 04 4 00 3 99 3 95 3 92 4 08
raquo 6 3 84 3 72 3 55 3 80 3 76 3 63
Babcock 3 92 3 93 3 77 3 93 3 80 3 83
Mojonn1er 4 09 408 4 08 4 10 413 4 10
